Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
DOCUMENt RESUME
ED 331 623 PS 019 488
AUTHOR Jackson, Jacguelyne FayeTITLE Multiple Caregiving aMong African AMericanS and
Infant Attachments: Issues and an ExploratozyStudy.
PUB DATE 90NOTE 65p.PUB TYPE Reports - Research/Technical (143)
EDRS PRICE MF01/PC03 Plus Postage.DESCRIPTORS *Attachment Behavior; Black Culture; Black Family;
*Blacks; *Child Caregivers; *Cultural Context;Cultural Differences; Exploratory Behavior; *Infants;Mothers; Sex Differences; Socialization
IDENTIFIERS *African Americans; Distress; Multiple Caregivers
ABSTRACT
Attachments of African-American infants should bestudied with a focus on cultural practices involved with multiplecaregiving. Assessments of African-American infant attachments thatuse standards of a culture in which care is provided by a primarycaretaker should be replaced by assessment based on the culturalperspective of African-Americans. This exploratory study tested 37African-American 1-year-olds in several separation/reunion situationsinvolving the mother and a second caregiver. Results showedconsistent reactions of infants to each attachment figure. There wasno bias toward mothers as a class. Results for various behaviorscales indicated: (1) infant use of attachment figures as a basis forexploration; (2) sociability with stranger independent of attachmentfigure presence; (3) distress when infant was left alone withstranger; and (4) no difference in seeking proximity, maintainingcontact, being inattentive, or resisting after each type ofseparation/reunion. Hypotheses that infants would respond similarlyto both attachment figures and use both attachment figures as a basisfor exploration were supported. The hypothesis that infants who werestressed would use caregivers for consolation was not supported. Itis inferred that culturally sanctioned patterns of caregiving do notproduce pathological relationships. Rather, they produce traitsconsonant with African-American socialization objectives. (BC)
***** 11111i*MailiMillift,01111111111*111111Mallaltaaltaallatifill*MIIMIIIIIIIMSPOtaReproductions supplied by EDRS are the best that can be made
from the original document...***************** ******* astassaa*********************assaaaaassaa,********
PO
u.O_011iORTONINt OF 164011101110Oe tooloosmeeDUCMIONAWMINCLINVOnanCd4
fERMIXilertus doeeud. teataluced as
Ism Ms mum Of OnlowshosOrquiltind st
Midi Id iiissov
Poodle! woo of osasonS Wad OHS &lawmost do 431 5IUSfd HIOsSiont ammo0111i Damen of waft
Multiple Caregiving Among African Americans and
Infant Attachments: Issues and an Exploratory Study
Jacquelyne Faye Jackson
University of California, Berkeley
Institute of Human Development
Data collection for the study discussed in this paper was
supported in part by a grant from Delta Sigma Theta, Inc. and a
research fellowship from the Center for the Study, Education and
Advancement of WomenUniversity of California, Berkeley under the
aegis of the Ford Foundation. Help in data collection and
analysis was provided byr Sharon Christal, Monte Cohen, James
Cole, Ethel Harrison, Sunjoo Kim, Christine Nakata, Frances
Nakata, Roxanne Smith, Verleeta White and Kim Bailey.
I gratefully acknowledge Diana Baumrind for her critical
reading and comments on this paper, and Wayne Shoumaker for his
CC) comments on data analyses. I also thank William Meredith for his
reading of a precursor to this paper and suggestions on
presentation of data analyses.
Running Head: MULTIPLE CAREGIVING AND AFRICAN AMERICANS
P1.14PERMISSION TO REPRODUCE THIS
MATERIAL HAS SEEN GRANTED SY
4504CieVA.khg. T63k4L"SamcksLIDv%
TO THE EDUCATIONAL RESOURCESINFORMATION CENTER tERIC)"
BEST COPY AVAILABLE 2
African Amer.
2
Abstract
Attachments of Afrielan Americans should be studied with a focus on
cultural practices ot multiple caregiving. Assessing African
American infant attachments using behavior standards of infants
from a culture with one primary caretaker is misplaced and should
be replaced by standards based on the cultural perspective of
African Americans. An exploratory study was undertaken to do
this. The monotropy thesis that all infants have a hierarchy of
attachments was tested and results form a substantial but
incomplete challenge to it.
African Amer.
3
Multiple Caregiving Among African Americans and
Infant Attachments: Issues and an Exploratory Study
African Americans are an overlooked cultural group who should
receive tho focused attention of scholars and investigators of
infant attachment and early social-emotional development.
Considerable discussion and controversies have been generated by
cross-national studies that raise questions about the validity of
the attachment paradigm and assessment procedures for diverse
cultural groups (e.g. Kuhn, 1910; Research and Clinical Center for
Child Development IRCCOl, 19901 Tavecchio & van ljaendoorn,
1987). However, there has been no corresponding attention to
African Americans as a cultural group within American society even
though they have distinctive cultural practices with implications
for the development of early attachments and they also subsfst in
an ecological niche that differs markedly from that of white
middle-class Americans. Nonetheless, African American infants and
young children have been incorporated in subject samples in
attachment studies in growing numbers with increasing frequency
(e.g. Lyons-Ruth, Connell, & Grunebaum, 1990; Crittenden, 1988;
Waters, Vaughn, & Egeland, 1980; Hock, 1980) as though the
analytic framework and assessment tools devoloped for white
middle-class Americans were valid for them. Specifically, a
salient aspect of African American infant and child care that
implicates attachment issues is multiple caregiving.
Traditional and contemporary African American child care and
ILA
African Amer,
4
socialisation at* based on thating of ostogiving by a mob**, of
parent figures irrespective of maternal marital status (Shimkin,
Louie & Prate 1978; Stack, 1970, 1974, 1975; Young, 1970). As
described by ethnographers (Shimkin et al., 1978, p.72),
"responsibility for protection, care, instruction, and discipline
of all children is diffused among related adults and, indeed, all
adults. Whether the relationships are actually centered in the
biological family depends upon circumstances and personalities."
Carogiving arrangements are closely associated with reciprocal
helping modes of life in extended family groups yet
distinguishable from family structure. Ceregiving transpires
across household, gender and age grov.ags of caregivers in the
network of kin and friends of the Pother. Focusing specificaliy
on infant and toddler aged children, Young (1970) described
typical mothers as important figures in their young children's
affective lives but nonetheless working women who relied on shared
caregiving arrangements to meet dual responsibilities of child
rearing and supplementing family finances.
Without acknowledging this aspect of child rearing as a facet
of African American culture, an irdirect mode ot addressing it has
evolved in the attachment research literature. For example. in
the early empirical literature on American infant-parent
relatiomhips explicit disadvantageous effects of multiple
caregivers were hypothesized for a study with a combined sample ,Lt
white and African American sub)Pcts and morg:nai supp;.,:t fc: +Pt.
Nzz..
African Ater.
hypothetic vat repotted (Caldwell, Rersher, Lipton, Richmond,
Stern, Eddy, Drachoan, & Rothman, 1963). Later, in a race
comparative study selection of subjects was restricTed to white
and (atypical) African American infants with mothers wka were
their unequivocal primary caregivers (Clarke-Stewart, 1973); less
favorable developmental outcomes were reported for African
American infants. Ultimately, African American infants were
studied using the strange situation procedure and evaluative
classification criteria which were derived from study of white
middle-class American infants with housewife mothers who were
their unequivocal primary caregivers (Bell, cited in Ainsworth,
Blew, Waters & Wall, 1976). A higher percentage of African
American infants than white middle-class infants were judged
insecurely attached to their mothers and the difference was
attributed to poor living conditions, including "... mother
absence from the home for long daily periods, and multiplicity and
discontinuity in regard to substitute caregivers" (Ainsworth, et
al., 1976, p. 166). This negative evaluation of multiple
caregivers stems from the infrastructure of the primary attachment
theorist's work (Bowlby, 1951, 1958, 1969, 1982) and two of his
main contentions: Diffusion of infant care is damaging, and
infants need continuous care from a principal mother figure
throughout tt.e litst years of life even though they may have
secondary caretakers.
There are cross-national. cross-cultural studies of i'lfan.
African Amer.
6
attachment in groups that practice various forms of multiple
caregiving (Leiderman & Leiderman, 1974; Marvin, VanDevender,
lwanaga, S. LeVine, & R. LeVine, 1977, Reed & Leiderman, 1981;
Fox, 1977; Tavecchio & van ljzendoorn, 1987; Smith & Noble, 1987;
Goosens, 1987), but their findings do not present a coherent
picture and their implications for attachments in African American
infants are unclear for methodological reasons (i.e. diverse study
conditions and criteria for assessing attachments). The most
recently reported cross-national, cross-cultural studies are
strictly limited to assessment procedures and evaluative criteria
derived from modal behaviors patterns of white middle-class
American infants with housewife mothers (e.g. Kermoian
Leiderman, 1986; Sagi, Lamb, Lewkowicz, Shoham, Dvir & Estes,
1985). In other words an etic perspective--using a paradigm from
one culture to examine another--(8erry, 1979) was a central
feature of the research. Some developmental psychologist
interested in cross-cultural study of attachment advocate the
continuation of an "imposed etic validity" approach (e.g. van
ljzendoorn, MO). In contrast, some cultural anthropologists
criticize the existing trend in cross-cultural attachment research
as overly reliant on evaluative criteria specific to traditicnal
white middle-class American culture (e.g. LeVine & Miller,
and characterize approaches to oross-ulturai study or t.ifant
development that are derived from psychoanalysis as prone to
Eurocentris (LeVine, 1q90,.
Afridan Amer.
7
Ainsworth's (1967) seminal field observational Study of
Ugandan infants was the first crossnational, cross-cultural work
on attachment, and is sometimes presumed to have implications for
Negroid racial groups because the subjects were black Africans.
Its implications for attachments in the cultural context of
African American infant and child care are limited, however,
because of differences between Ganda and African American cultures
with respect to the shared caregiving factor; the Ganda resemble
white middle-class Americans more than African Americans.
Ainsworth's Ugandan subjects were from the relatively Westernized
ethnic Ganda for whom a nuclear family removed from an extended
family social context was the modal family form (c.f. Nahemow,
1964), and who did not practice multiple caregiving extensively
(Ainsworth, 1967). Moreover, as might be expected in the Ganda
cultural context Ainsworth found that the extent to which a Ganda
mother was a principal caretaker of her infant was a salient
discriminator between what she judged to be qualitatively
different attachments: a) securely attached infants received most
t their care from their mothers, b) insecurely attached infants
received a substantial amount of their care frum individuals other
than their mothers, and c) unattached infants received little car.?
AltogPther tAinsworth, 1967 12.91, p.395-396).
A kev 1Ssue and ditticnity in attempting to tap existing
theory and research to understand early attachments in African
Amiricans is Bowlby's (1:469, 1962) unchallenfe,.; contentic,n
African Amer.
8
contemporary cultural variations ate superfluous to core elements
of human attachment propensities. According to his explication
attachment capacities are biologically determined and were shaped
by evolutionary processes culminating in the social context and
subsistence ecology of human life in prehistoric hunter-gatherer
societies. Pursuant to this, attachment adherents have focused
on the contemporary tKung San (Lee, 1979; Lee & DeVore, 1976;
DeVore 6 Konner, 1974) of the South African Kalahari desert who
are one the few remaining huntergatherer peoples as a model of
primordial human infant-mother attachments. Among the San the
mother is clearly the primary caretaker of the infant in the first
years of life. However, very recent ethnographic reports of
infant care among the Efe of the Ituri Forest of Zaire 'Winn.
Tronick & Morelli, 1989; Tronick, Morelli & Winn, 1387; Tronick,
Winn & Morelli, 1985) give reason to question singular reliance on
reports of the !Kung San as a model for infant care in prehistoric
hunter-gatherer groups. The Lie are a seminomadic group
traditionally considered a hunter-gatherer people (Winn, Tronick 6
Morelli, 1989) who practice extensive multiple caregiving of
infants from birth and throughout the first years of life. (To
date, there are no reports of analyses of Efe infant-caregiver
attachments.)
In looking tor c-cnfempotary expressionsc of adaptive fcres
presumably evolvel infant attachment cafacities, attachment
teseatchett: have studiel aFf.'endant mayorities witn:n 7ourtriv-:
African Met.
9
that dominate the global political economy that also have
sociocultural niches that diverge from prehistoric hunter-
gatherers in innumerable ways (i.e. the middle class of the United
State, Germany, Japan and Israel). Even in this special subset of
world peoples no common modal pattern of infant-parent figure
attachments has been found. Nonetheless, current discussion of
appropriate criteria for judging adaptive, psychologically healthy
infant-caregiver attachments revolve around modal behavior
patterns for traditionally reared white middle-class American
infants (e.g. Kuhn, 1990; RCCCD, 1990), and only indirectly
acknowledges the possibility that characteristics of attachments
may inevitably reflect culturally determined effects of
socialization to a greater degree than universal biological
propensities.
A diiferent approach to examining early attachmentp in
African Americans is needed. An emic approach--one that frames
analyses from an African American value perspective--(Berry, 1979)
is one alternative adopted for the study to be reported. An emic
approach recognizes that what is adaptive for African Americans
may not be identical to what is adaptive for white middle-clasi
Americans (Ogbu, 1981). An emic approach also dictates a
recognition of African American sociocultural structure as it
pertains to child care, how this structure relates to group
subsistence resources and constraints, and what African Americans_
value in relationships that are social-emotionas outcomes 1
,47.4
AfriCan Amer.
10
caregiving efforts. With respect to this last point, African
American scholars (e.g. White, 19871 farness 1980) maintain that
strength and extensiveness of social relationships with.n African
American kin and codounity groups constitute the essence of group
identity and a critical element of personal identity. Tronick and
his associates (Tronick, et al., 1987) make a similar point
concerning the Efei they maintain that the practice of shared
caregiving is a major factor shaping group identification during
Efe development.
e_and_
American Infamt Cate
Constraints on striving for economic self-sufficiency and
self-control have been the foremost challenge to adaptation for
African Americans throughout their history in the Americas and
continue to be. A primary, group strategy for addressing this
challenge has been paid workforce participation for women with
children. Sixty-six percent of married African American women
with a child under three years old work and 44% of single African
American women with children that age work (Hayghe, 1986;
Children'i Defense Fund, 1985). Regardless of marital status,
African American female labor force participation is pot-itively
correlated with education and family economic status, and African
American females who are high school and college graduates work
much higher rates than their white counterparts (Meisenheimer,
Emploved Atrican American mothers of young children rirt-
11
African Amer.
11
also distinguished by reliance on rela:ives who are extended
family members to care for their children while they work
(O'Connell & Rogers, 1983, pp.24-28). Moreover, the pattern of
African American maternal employment has existed throughout this
century (U.S. Bureau of the Census, 1974, p.338) and therefore
distinguishes African Americans from majority Americane who have
only recently experienced large numbers of working women with
small children in the work force.
One explanation for this high rate of maternal employment is
current end historical discrimination and disadvantage in
employment for both male and female AM:I.:. Americans. Even at
the highest education levels African American men earn less than
their white counterparts and suffer higher rates of unemployment
(Meisenheimer, 1990). This situation has made maternal income a
vital element of financial viability in African American families.
Moreover, African American women and men have worked and continue
to work under conditions that are unattractive to others such as
night, evening and split shifts (Presser, 1987, 1986) and second
jobs outside of the home (Stinson, 1990).
African American adults have adopted flexible family roles in
part to accommodate the de-iands of employment. In two-parent
families spousal roles are egalitariam and one concomitant 16
routine Involvement of fathers in caregiving tasks (e.g.
Billingsley, 1968; Scanzohi, 197b; Beckett, 1976). Relatives el
single mothers are alsc caregivers when mothets work, ev(-r if $..3
African Amer,
12
relatives work themselves (Presser, 19893, host likely, the
cultural rihic of affective as well as material reciprocity that
generates the relatioships between adults that produce these roles
(Stack, 1974) creates an affective climate for the caregiver-child
relationship that differs from that of the hired caregiverchild
relationship in majority American culture (Werner, 1984). To
complement this caregiving context infants and children of
employed African American mothers are socialized to respond in a
positive, equitable way to caregivers who are essentially
interchangeable, and a generally gregarious personality is
cultivated (Young, 1970). The question about child psychological
development that ensues from this socialization thrust is, do
African American infants develop multiple relationships that are
essentially equivalent attachments? To begin to address this
question in an exploratory way, the findings of a study of healthy
African American infants with working mothers will be reported.
Tho Study Questions in Attachment Terms
In attachment terms the focal question is, do African
American infants develop monotropic bias toward one caregiver and
demonstrate a stronger tie to one person from their set of
caregivers? As delineated by Bowlby (1969, 198Z) the monotr3ry
principle specifies that an infant with nonpathological sccio-
emotional development hill develop an attachment to a preferred.
primary mother figure even though he or she may have several
attachment figures dnd will show a clear bias by dire-t:ng
African Amer.
13
stronger attachment behaviors toward the preferred figure. In
addition, this bias is most evident when focused attachments are
consolidated which occurs when infants can fully locomote at about
one year old. Ainsworth (1982) contended that monotropic biac
could not be detected from naturalistic observation of infant-
caregiver interaction because stress activated attachment
behaviors would be unlikely to occur in low stress, familiar
settings. She (Ainsworth, 1982) maintained that monotropic bias
would surface when the infant was moderately distressed, that is
when tired or ill. This suggests that an induced state of infant
distress like that presumed to occur in separation-reunion
laboratory studies of attachment would provide a sufficient
precondition for a test of the monotropy principle for African
American infants.
Another question is, which behaviors are valid and
arpropriate critwrtu for detecting attachments? Two contrasting
affective states must Le induced in an infant before true indices
of attachment can be discerned. First, there are effects of a
distressed state which activate what is referred to in a delimited
sense as the attachment behaviorai system. Bowlby (1958)
originally identified infant behaviors that promote close
proximity to an attachment figure as expressions ot an activated
attachment behavioral System. Second, there are ettects of a
contented ana composed state which promote activation of the
exploratory behavicra, system. Adding to rowlby's r(,cur
1.;
African Amer.
14
proximity seeking, Ainsworth (1967) identified an infant's use of
the attachment figure as a secure base for exploration of the
physical environment as an expression of an activated exploratory
behavioral system. Subsequently, Bowlby (1989, 1982) and
Ainsworth (1969, 1978) added separation distress and variation in
response during reunion with an attachment figure following a
separation as indicators of attachment. Currently, reunion
behaviors are emphasized because research has shown them to be
important in predicting different developmental trajectories for
infants from Euro-american cultural groups. However, the
significance of reunion behaviors for infants from other cultural
backgrounds remains unclear. Given the paucity of attachment
research on African Americans, it seemed reasonable to use the
full range of behaviors as appropriate criteria for evaluating
attachments in African American infants.
Concomitant to selection of behavioral criteria tor assessing
attachments is the question of interpretive validity of responses
presumed to be indices or attachment. Results of a pilot study
(Hansen, 1980) i'ndicated the importance of this issue for
evaluating laboratory Gehavior of African American infants. 1hk
object of study was behavior of African American infants with
working mothers from low risk backgrounds in strang0 situation
sessions. Even though the strange situation is presumed to be
milaly stresstul to infants, the majority ot plIct study intantF
dtd not appeal Etressd Gy the ,zepariitic:nt, entitt,i-d in strAnge
African Amer.
15
situation sessions. Instead, playfulness and exploratory
behaviors were salient, even in reunion episodes. Superficially,
infant responses fit the profile of white middle-class American
infants judged avoidant and insecure in their attachments on the
basis of Ainsworth's t Ainsworth et al., 1978) classification
scheme. However, viewed more closely pilot study infants appeared
outgoing and casually inattentive due to preoccupation with play.
The pilot study implicated the need for modification in the
design of laboratory procedures chosen to study attachments in
African American infants as well as methods of infant behavior
analysis. Modified procedures need to surpass the strange
situation procedures in detecting the specific nature of infant
affective reponses to separations, reunions, and to the laboratory
setting in general. In other words, laboratory and assessmint
procedures need to make it possible to determine whether or not
attachment, exploratory or both behavioral systems are in fact
activated during a session. An experimental procedure
(Kotelchuck, 1976) was chosen for tne study to be reported that
held promise of meeting these needs. It has the added advantage
of allowing direct comparison of infant resporses toward two
attachment figures on all of the attachment behaviors of interest
based on infant behavior in a single laboratory session.
Research hypotheses are tramed on the basis ot an emic
perspective as well as attachment concets. The premtse is that.
African American cultural practices c.f srldrel cAregiving ar.1
African Amer.
18
Compatible with normative infant responses that do not indicate
incipient psychopathology. The first hypothesis is that African
American infants with multiple caregivers use their attaohment
figures as secure bases for exploring the environment when
composed and also seek proximity to them when stressed. Practical
constraints of the laboratory procedure limited this exploratory
study to comparison of infant responses to onl!, two of their
multiple caregivers. The mother and tha second most impPitant
caregiver according to her designation were the attachment figur.A1
chosen. Framed on the basis of this constraint, the second
hypothesis is a direct challenge to the monotropy thesis: African
American infants show a pattern of similar strength of responses
toward two of their attachment figures for a full range of
attachment related behaviors, and their two attachment figures
serve the functions of a primary attachment figure equally well in
a laboratory setting. In operational terms the hypotheses are
that normative responses of African American infants reveal that:
a) infants' responses to their attachment figures are similar And
equitable for a full range of attachment related behaviors, b)
experimental manipulations of attachment figures' presence 7ause
infants, it composed, to use their attachment figures as se:ure
bases for exploring the environment. and c) iranipulations of
attachment figures' presehoe .7ause infants, if stressed, to use
their attachMent figures as sour(7efi consolati.ln. Support for
a.i !Ps i tht...,e hypt,!.eses is netessars, tc, fully refute
monotropy thesis.
Method
African Amer.
17
Eutiects
Twenty-one male and 16 female one-rpar-olds along with their
mothers and second attachment figures participated in the study.
The Jrifants were between 12.1 and 13.5 months old at the time of
the laboratory session. Local birth records were used to identify
and recruit infants with the following characteristics! (a) no
birth defects, (b) a mother 20 years old or older, (c) both a
mother and a father who were African Americans, and (0 a mother
who had an occupation or was a student. Each mother was asked to
identify a second adult who was the most important caregiver in
her infant's life other than herself. As a result of this
request, all 32 of the biological fathers in two-parent families
participated and five female relatives participated with the five
single mothers in the sample; there were 37 second attachment
figures in total. Occupational and educational information
revealed that working and middle classes were equally represented
among the famiiies (Riau & Duncan, 19C7).
Interviews with adult participants revealed that for both
middle 3nd working class families, caregiving arrangements wre
diverse and complex because parents' work schedules were
incongruent and frequently entaileo unconventional hours (e.g.
attributable to evening or split shifts, demanas of a professioral
occupation, second joPs, etc.); 7E,Gli ct muthers and tathl?rs 11J r.,t
African Amer.
work 9to-5 schedulee, and in 65% of the twoparent families
parents' work schedules were incongruent. Thirty-one of the 37
infants received routine supplementary care from one or more
caregivers acquired through parental kin and friend networks in
.dition to care from the attachment figures participating in the
laboratory session. For all but one of the 37 infants,
supplementary care was provided in private home settings or in day
care homes; the exceptional infant attended a day care center.
Procedure
The Kotelchuck (1976) procedure consists of 13 thrtie-minute
episodes, each of which functions as a separate experimental trial
(i.e. treatment condition). It is compatible with principles of
design for time-series experiments (Campbell & Stanley, 1966;
Kratochwill 1978), which makes sequential changes due to treatment
effects detectable. The presence of each attachment figure and the
presence of the stranger are three independent variables that are
coordinated and experimentally manipulated to produce cyclical
patterns of exits and reentries of adults while the infant remains
in the room. The order of attachment figures' initial exits and
reentries was counterbalanced as indicated in table 1,
Insert Table 1 about mere
which als)
shows which adults were present in thp room with the infant durIrt!
each epl!-,ode.
African Amer.
19
There is a distinctive structural feature of the KotelChuck
procedure of central importance for the study. As indicated in
table 1, infant participants experienced two different types of
separations from their attachment figures which were not
contiglious. in one type, one attachment figure departed the room
but the second attachment figure remained with the infant (i.e.
episodes 2 and 8), and minimal or no separation stress reactions
were expected. In the other type, only one attachment figure was
in the room at the start of a separation, and the departure of
that attachment figure left the infant alone with the stranger in
a completely unfamiliar setting (i.e. episodes 4 and 10), and
separation stress was expected. This difference made it possible
to compare the effects of the two types of separations to
determine if the second type actually elicited attachment
Aaviors of the sort that ensue from a stressed state (Bowlby
1969, 1982). The logic lot the comparison is based on Bowlby's
(1973) observation that the presence of familiar people markedly
diminishes or Precludes an infant's distress when separated from
an attachment figure as well as withdrawn and angry behaviors when
reunited with the attachment figure.
The procedure was initiated when an intant and both
attachment figures were shown into a 12.5 ft t3.81 mi by 16 ft
(4.66 m) playroom at a university facility. The layout of
furnishings provided a rug with age appropriate toys in one
quadrant along with three dduit sized chairs anm two chi.d
African Amer.
20
chairs near the outer perimeter of two sides of the rug.
Attachment figures were directed to be seated and were ;nstructed
to allow the infant to explore the room and toys independently in
the first episode of the session. However, attachment figures
were encouraged to respond to their child in whatever manner
seemed natural to them in all other episodes of the session. The
entire session was videotaped through a one-way view mirror by a
camera placed in an adjoining room.
The stranger was always a young African American female adult
college student. Her role was designed to make her a
nonthreatening, attractive social novelty when the infant
initially encountered her, friendly at all times, and to encourage
sociable infant responses. However, two potentially disconcerting
events were built into the stranger role to challenge infant
c- posure and self-assurance during the episodes when the infant
and stranger were alone. During the first episode when the infant
and stranger were alone the stranger cffered the infant a clear
plastic container of visible cookies that could not be openea; it
was assumed that' the frustration of not being able to get the
cookies would be mildly distressing. During the secona episode
when the infant and stranger were alone the stranger found
occasion to burst a balloon, ostensibly as an accident; the
assumption was that the sudden noise of the explosion would be
upsetting.
bevenOrnt
African Amer.
21
There were seven dependent measures, all of which were 7-
point ordinal scales for rating infant behavior in response to an
adult or in the presence of an adult (Jackson, in press). The
first three scales were developed for the study on the basis of
pilot study (Hansen 1980) outcomes. These scales and their high
and low anchor points are: (a) aloloratim-exploratory activity
in the presence of an adult without an overt social response; from
extensive, independently initiated physical mobility away from
adults and focused attention on a specific item, to an absence of
exploratory or play behavior independent of social interaction,
(b) pociabilitv-sociability with the stranger; from infant
initiated and sustained interaction with the stranger or continued
interaction with the stranger in spite of minor disconcerting
events associated with her, to an absence of attraction to Or
positive social response to the stranger for the whole episode,
and ic) distress-visible distress reactions to an attachment
figure's departure from the room; from loud crying and other
conspicuous protest behaviors that either could not be alleviated
while an attachment figure was absent or aborted the attachment
figure's efforts to leave, to an absence of distress behavior or
one occurrence of less than 5 seconds.
The four additional scales were modeled after the interactive
behavior rating scales developed by Ainsworth Ainsworth et al..
1976), although the name assigned to one scale differs from its
Ainsworth counterpart. The rating cri'.eria tor ail four scaie
African Amer.
22
are the same as those of the Ainsworth scales except for minor
modifications such as the addition of explicit time parameters for
making some judgments. The four scales and their high and low
anchor points are; (a) proximity seekint-seeking physical
proximity to a reentering attachment figure; from the combination
of infant initiative to attain contact with an attachment figure
and actually achieving contact through that initiative, to an
absence of infant initiative to gain contact or proximity to the
attachment figure, (b) =Aka_ maintaininc-maintaining physical
contact with a reentering attachment figure once contact is
attained; from 90 seconds of contact sustained by the infant's
efforts or 2 minutes of contact fully accepted by the infant, to
no contact or contact of less than 5 seconds which the infant made
no effort to sustain, (c) j,mattentila (similar to Ainsworth's
avoidancel-not attending to a reentering attachment figure; from
ignoring a reentering attachment figure even though the adult made
several efforts to get the infant's attention or turning and
looking away while being picked up by a reentering attachment
figure, to immediately acknowledging a reentering attachment
figure and continued acknowledgement throughout the episode, and
(d) resistinivresisting physical contact with an attachment figure
or de:tart and aggressive behavior toward an attachment figure;
from aggressive resistance during contacts with an attachment
figure, to an absence of defiance, aggression and resistance or
physical contact with ap ,Ittachment
African Amor.
23
Six trained raters and the investigator analyzed the
videotapes to generate ratings on the seven behavioral scales.
Each tape was rated by two raters and the scores assigned were the
average of the two raters' judgments for all scales and for every
tape. Interrater reliabilities were obtained from ratings of a
subsample of five tapes and ranged from 88% to 100% agreement for
the seven scales.
Destins for Data AnAivses
Two interrelated sets of analyses were planned to address the
study hypotheses. The first set addressed the questions of order
effects of the procedure and similarity of infant responses to the
two attachment figures. For these analyses anly ratings from
episodes associated with the influence of a single attachment
figure were used. Because the order of attachment figures'
departures was counterbalanced for the sa,fle as a whole, the
specific episodes associated with a mother or second attachment
figure would depend on which attachment figure departed first in a
particular case. The first set of analyses was also a technical
pre:equisite for the second set: if infants were found to be
responding to their two attachment figures similarly, then
analyses to detect experimental effects could proceed as though
attachment figures were equivalent independent variables for
purposes of experimental manipulations.
The second set of analyses addressed the question of whether
Cr not attachment hna evploratory behaviors were indu,:ei in t?',e
African Aber.
24
sample as a whol by manipulating the presence of the two
attachment figures and the stranger. For all of these analyses
each episode of the procedure was treated as a separate trial, and
infant behavior in selected episode'', was examined with planned
comparisons of differing structures for each of the dependent
variables.
Data used in the second set of analyses were converted from
ratings on the original scales of measurement to rank data,
because descriptive statistics for the rating data of episodes to
be compared revealed that the assumptions of repeated measures
analysis of variance were not met. The variances were unequal for
six variables, paired episode correlation coefficients were
unequal for the seventh variable (sociability), and for some
variables such as proximity seeking both variances and paired
episode correlation coefficients were unequal. Friedman's
nonparametric procedures for within subject ranking of data
(Marascuilo It McSweeney, 1977) were used as alternatives to
analysis of variance for repeated measures procedures. Each
subject's ratings on a dependent measure were ordered and
corresponding ranks were assigned. Then the converted data for
each dependent variable was treated in the following way: (a)
rank values were added across subjects for each episode involved
in a planned comparison, (b) a cuefficient was assigned to the
mean of the summarized rank values for each efisode to make two
counterbalanced sets of episodes for a comparison. and ,c1 the
African Amer.
25
statistical analysis was finalized as a confidence interval. Even
though they are invalid, results of comparisons based on analysis
of variance for repeated measures calculations are also reported
for confidence intervals that were not statistically significant
based on Friedman methods to enhance the discussion of valid
results.
In addition, planned comparisons with nonparametric methods
were used to analyze data for the girls' and the boys' subsamples
as well as for the whole sample, because gender differences were
examined as an aspect of normative infant behavior. Post hoc,
intercorrelations of summary scores for each dependent variable in
the original scale of measurement were examined to explore the
possibility that individual differences accounted for the initial
observation of marked differences in variances for episodes used
in the planned comparisons.
Results
Comparison of Responses to Attachment Figures
To test for departure order effects, two-sample t-tests were
performed for all variables in episodes where a rating was
attributable to the influence of only one attachment figure. This
was all episodes except one, seven and thirteen. No order effects
were found; Only two of the thirty-four tests wvie significant
which is what one would expect to find by chance.
Infants' behaviors influenced by mothers alone correlated
with behaviors influenced by second attachment figures alone
African Amer.
26
revealed consistent similarity of reaction to the two attachment
figures across variables. Table 2 reports the results of the
Insert Table 2 about here
seven tests. For six of the seven variable pairs examined,
Pearson product moment correlations were statistically
significant. The absence of a significant correlation between
resisting influenced by mother and resisting influenced by the
second attachment figure was the only exception to the pattern of
correlations. This may be due to the fact that the overwhelming
majority of ratings for resisting were at the lowest scale point,
indicating an absence of resisting behavior.
A second group of analyses addressed the question of whether
of not the infants were biased in their attachment responses
toward mothers as a class. The data used for these analyses were
based on difference scores obtained by subtracting ratings for
behavior influenced by the second attachment figure (i.e. "adult
IQ") for a variable from behavior influenced by the mother for the
same variahle for each subject in the sample. Matched pair t-
tests for all of the Opendent variables revealed no statistizall
si.'niticant bias toward mot:sets for any of the seven variable
pairs examined, or no '-vidence ot pretelence for mothers. Tt-e
results of this group of tests are summarized in tat,1 3.
African Amer.
27
Insert Table 3 about here
Experimentki__Manioulations
This second set of analyses detected whether or not
manipulations of attachment figures' and the stranger's presence
were effective in inducing changes across episodes in infants'
exploratory and attachment behaviors.
Pooloration. Changes which would reveal infant use of both
attachment figures as secure bases for exploring the physical
environment were predicted tor this variable. Specifically, the
infants' exploratory behavior would shift froa high levels when
only attachment figures were present (episodes 1,2,6,7,8,12 and
13) to low levels when only the stranger was present (episodes 4
and 10), and achieve intermediate levels when one attachment
figure and the stranger were precent (episodes 3,5,,8 and Il).
The expectation was that when the stranger and an attachment
figure were present the infant would vacillate between social and
nonsocial forms of exploration, or be wary of the stranger and
therefore distracted from exploration. Figure 1 presents the
Insert Figure 1 about here
curvilinear pattern delineated by the means of ranks for the
exploration variable. The results of the three contrasts
revealed in tables 4 and 5 show that the predicted changes
4) 4'
African Amer.
28
Insert Tables 4 and 5 about here
actually occurred; infants did use their attachment figures as
bases for exploring the physical environment.
Sociability. Normative reactions to the stranger were not
predicted, and therefore directional effects were not hypothesized
for the planned comparison. There were three possibilities;
infants would be more wary and therefore less sociable when aime
with the stranger than when an attachment figure was present, they
would be more sociable when alone with the stranger, or they would
be equally sociable in all episodes where the stranger was
present. Statistical rerults in tables 4 and 5 show that level of
sociabil!ty was the same in all episodes where the stranger was
present, indicating there was no change in response to the
stranger attributable to the presence or absence of an attachment
figure and the stranger's behavior.
Pietrell. Less distress was predicted for episodes in which
one attachment figure remained in the room to reassure an infant
following the other attachment figure's departure (episodes 2 and
8), than for episodes In which an infant was lett alone with the
stranger (episodes 4 and 10). The comparison results revealed
%see tables 4 and 5) that n average infants did display more
distress when lett alone with the stranger than when lett with the
second attachment figure.
African Amer.
29
proximity Seeking,. The prediction was that infants would be
more active in seeking out a returning attachment figure after
being alone with the stranger (episodes 5 and 11 when some stress
was likely), than in seeking out a returning attachment figure
rejoining the infant and an attachment figure who stayed in the
room (episodes 7 and 13 when stress was unlikely). A contrast in
level of proximity seeking consonant with the prediction would
indicate attachment behavior ensuing from both a stressed state
and effectively uninhibited attachments. Statistical results
revealed no difference in proximity seeking behavior in the two
types of reunions. For the sample overall, this indicated absence
of a stressed state, which would prec-.r. proximity seeking as an
attachment behavioral response, absence of uninhibited attachments
or both.
Contact Maintaining. The prediction was that infants would
engage in more contact maintaining behavior with an attachment
figure in the reunion episodes following being alone with the
stranger (episodes 5 and 11 when infants were likely to be
stressed), than in episodes where a returning attachment figure's
reentry had been preceded by the presence of one at..achment
in the loom with the infant (episodes b.7.8.1L and 13 when intart:
were unlikely to be stressed). A difference in extent of contat
maintaining ,7onsistent with the prediction wi)%Jid indicate
attachment behavior stemming from a combination of a stressed
state ord atfectlif,lv attzichment,.
African Amer.
30
Contact maintaining is an attachment behavior similar to
proximity seeking except that the duration of contact indicates
strength of attachment response, while promptness of proximity
seeking indicates strength of attachment response. To account for
the difference in how these attachment behaviors are gauged, the
planned comparison for infants' contact maintaining behavior
involved data from seven episodes of the session (see table 4).
As shown in table 5, the comparison was not statistically
significant, which indicated that on average infants' contact
maintaining behavior toward an attachment figure was no greater
just after having been alone with the stranger than after having
been in the room with an attachment figure.
inattention. Infants' inattention to reentering attachment
figures was exam!ned in a comparison with the same structure as
the one for proximity seeking, but no direction of effects was
predicted. Three normative outcomes were possible. First,
infants could show more inattentive behavior toward a reentering
attachment figure in reunions after the stranger and infant hao
been alone (episodes and 11) than in reunions atter the infant
had been in the room with one atL4chment figure (episodes 7 ano
13), if intants were both stresse.: and had attectively inhibited
relationships to their attachment figures. Second, infants coulc
st.ow iess inattentive r-enavior toward the reentering figure in
reuni,:ris atter the stranger and infant nad Leen alone tepisoi s
and 1:: th:sn in reunIcns atter tt,e ihfant i4nd cnt attacrmvnt
African Amer.
31
figure had been in the room together (episodes 7 and 13), if
infants were both stressed and had uninhibited relationships to
their attachment figures. Last, there could be no difference in
level of inattention in the contrasted episodes becausl: of lack of
treatment effectF of the contrasted reunion situations. Results
revealed no difference in inattention in the two types of
reunions, or an absence of treatment effects for the sample as a
whole.
Resistink. The planned comparison for resisting behavior had
the same format as the one for contact maintaining for the same
reason. However, no direction of effects web predicted. The
alternatives for normative outcome were: a) more resisting
behavior toward a reentering attachment figure in reunions after
the stranger and infant had been alone (episodes 5 and 11) than in
episodes after the infant had been in the room with an attachment
figure (episodes 6,7,8,12 and 13), b) less resisting behavior
toward a reentering attachment figure in reunions after the
stranger and infant had been alone (episodes 5 and 11) than in
episodes after the infant had been in the room with an attachment
figure +episodes 6,7,6,1Z ani 13), cr c) no difference in level
resisting in the contrasted episodes due to lack 4 treatment
effects for the resisting variable. Thit. last (4 the possible
outcomes was tt.e onv abtathea; there wols no nnmatIvP resisting
response.
ReVerlted AltY,)1C.
African Amer.
32
technically invalid, confidence intervals tor sociability,
proximity seeking, contact maintaining, inattention, and contact
resisting are presented in table 6, based on analysis of variance
Insert Table 6 about here
for repeated measures methods of calculation (Marascuilo, 1971. )
First, the confidence interval for proximity seeking gave the
appearance of statistical significance. It is spurious because
two assumptions for the methods of calculation employed are not
met: Paired episode correlation coefficients are unequal, ranging
from rage7depligi .03 to Leesowe .56, and the episode variances
are moderately unequal as shown in table 4. Next, table 6 shows
that the confidence interval for contact maintaining also appears
to be statistically significant. This is a spurious finding as
well because OIL ..equisite assumption of equal variances of the
episodes compare not met; table 4 shows the markedly unequal
variances. Hc.e. . if the results for these two variables had
been valid they .-.1c1 have provided support for the research
hypotheses.
Gender Comparisons. Analyses cf data for the boys' and
girls' subsampies used the same set of variables and the same
structures LA 1.ianned c6mparison; applied to the whole sample
data. Only ohe sex difference was fnund, isys and girls differe.:
pat*.efnF, of 0.strest. cf-!,phnst.. Fr t,ovs thetp
African Amer.
33
significant diffarence in extent of distress in response to the
absence of both attachment figures as compared to the absence of
only one attachment figure. Girls, however, were more distressed
by the absence of both attachment figures than by the absence of
only one. This discrepancy is reflected in the statistically non-
significant comparison for boys and its associated one-tailedCaer 41:\ ...,.,,,,,.,................. - .. ... ...a ...." Pi A
. -- confidence inter:#:12 -.53 1.4, Var. Y4p a .32 (critical value_.---
?5'ali act 1, RA e vi 0 ....----A--i---,-I-7z----------._......... ..... *a - 1.00.4......."."4°.
based of..1;j, a 6.25, 01 a 3, a ( .05), in contrast to the
....------* '
C.)";" Stiallie statistically significant comparison for girls and its relatedA 05 44 pgrscrobr
one-tailed confidence interval corrected for tied ranks c/.- 4° 141 sc
-1.87 1.63, Var. o: .42 (critical value based on 14.1.62 6.25,gar-,
di 3, a< .05). For the boys' comparison (ms21) means of the
ranks and corresponding weights for episodes 2 and 8 were 2.02 (1)
and 2.71 (1) whereas means of the ranks and corresponding weights
for episodes 4 and 10 were 2.45 (-1) and 2.81 (-1). In the girls'
comparison (n1=16) moans of ranks and corresponding weights for
episodes 2 and 8 were 1.88 (1) and 2.19 (1), but means of tanks
and corresponding weights for episodes 4 and 10 were 2.88 --1) and
3.06 (-1).
Intercorrelations of dependent variables
For each 0 the seven dependent measures ana tor each
subject, a score was generated that was the average of all eplsz.:
ratings in the originai scale of measurement. lntercorrelations
of variables based on tt.ose composite average scores produred twO
cIuEt.-rt:. In t!-A. !list 1.:ucter itOXIMItv
African Amer.
34
maintaining and distress were positively intercorrelated. In the
second cluster exploration, and sociability were negatively
correlated with proximity seeking, contact maintaining and
distress while inattention was negatively correlated with
proximity seeking and contact maintaining only. There was one
significant correlation outside of the two clusters and one
correlation that approached statistical significance; sociability
was positively correlated with inattention and exploration.
Table 7 presents the correlation matrix for these results.
Insert Table 7 about here
Discussion
The first hypothesis stated that the study infants would
demonstrate a pattern of similar responses toward their twc
attachment figures across a full range of behaviors that pertain
to an attachment relationship. The results of the correlation
tests and the tests for mother bias gave clear support to this
hypothesis, and constitute the first element of a challenge to the
view that all one year old infants with multiple attachments are
monotropicaily biased toward one superordinate attachment figure.
Ooteichuck (1976) used the same laboratory procedure with
t.raditic:nally reared ::-month-oid white Americar intants and tou !
their responses biased toward mothers even though infant response
to m.-)theri and tathers Lr imilar. !*(".k Wsr0 t!le Salle
:4.1
African Amer.
35
procedure and, like Kotelchuck, found Israeli kibbutz reared
infants of this age biased in their reunion reiponses toward their
mothers as compared to their metaplot who were their primary
caretakers. The pattern of unbiased, similar responses of our
subjects to their two attachment figures stands in marked contrast
to these other groups. Most probably, our findings are
attributable to specific features of the African American cultural
and social background of our subjects.
African American culture (Shimkin et al., 1978; Stack 1974)
assumes that all famWar caretakers of a young child, including
fathers and other males, are capable of performing both the
affective and practical functions that are theoretically
attributed to attachment figures. In an ethnographic study, Young
(1970) reported that African American toddlers raised in
traditional settings are equally responsive to most of their
caretakers yet develop and retain relationships to their mothers
that are affectively secure. The results of tests of the first
hypothesis suggest that the mother and at least one other major
caretaker might actually be equally effective as elicitors of
nonpathological attachment responses, and provided a technicaW
defensible starting point tor making such a determination.
Results provided support for our second hypothesis as well.
The prediction was 'hat infants would use bce.h dttachment
ds bases for exploring the environment, and the findings ffIVPale-:!
that they definitely lid sn in t?xplotin4 the !.!-Iy. 1.:0;
3i;
African Amer.
36
of the laboratory playroom. In this respect, infants demonstrated
behaviors indicative of adaptive infant-attachment figure
relationships according'to one of Ainsworth's (Ainsworth et al.,
1978) and Bowlby's (1982) pivotal criteria, and these behaviors
were not attributable to tangential factors such as the presence
of the stranger.
The primary purpose of the sociability contrast was to tease
apart the impact of attachment figure absence from the stranger's
presence and behavior during the two episodes when the infant and
stranger were alone, for purposes of interpreting expected change
in infant exploratory behavior. The fact that there was no change
in infant sociability toward the strange when attachment figures
left the room as compared to when they were present, indicated
that the drop ln exploratory behavior occurred as a response to
attachment figures' absence only; the stranger's presence had no
impact on this behavior. Concomitant to this, no impact of
stranger behaviors designed to be mildly disconcerting were
observed either.
The third hypothesis stated that the infants would use their
attachment figures as sources of consolation, if stressed.
'2upport for this hypothesis was not found. lendering a decisivt .
test of the monutropy thesis incomplete. However, the
cuntiguration ot Fertlnent reiiults highlirit.: the importance or
giving attention to the conditional nat..ite ut the hypothesis ac
well tne cOnditiuhal nature 61 attachrif-n t!e-ry reCircl!'V t
3 7
African Amer.
37
circumstances under which infants will protest separation from an
attachment figure and/or seek proximity to and physical contact
with an attachment figure. These findings also highlight the
value of built-in procedural checks to confirm that the attachment
and exploratory behavioral systems are actually activated for
infant participants in laboratory sessions where they are presumed
and assessed.
The boys did not demonstrate more distress when left alone in
the room with the stranger than when left with an attachment
figure, even though the girls did. Neither boys nor girls
demonstrated less sociability toward the stranger when left in the
room alone with her than when they encountered her in the presence
of an attachment figure. In this connection, it is important to
recall that an unchanging level of sociability was found oven
though there were built-in elements of stranger behavior in the
absence of attachment figures designed to challenge infant
composure. Togethe. the results of the planned comparisons for
distress and sociability behavior indicate that the infants were
minimally distressed during separations from both attachment
figurec, and suggest that most of them were not stressed at !he
point ot reunion with attachment figuies after having been alore
with the stranger.
The planned comparts_ns fur infant reunion behaviors, tIc't!.
those Indicative of uninhibited attachment and those Indicative :1
aitectivaly duiensiv*, ottachment, were structural t ) ds;.orfolt
African Amer.
38
role of a stressed state in the infants' reunion responses as well
as identify types of predominate attachment responses. Reunion
behaviors under conditions when infant stress was improbable were
compared to reunion behaviors when infant stress was probable
(i.e. following episodes of having been alone with the stranger).
Statistically vaiid results showed that infants' proximity seeking
and contact maintaining behaviors, which indicate uninhibited
attachment, were the same in the contrasting conditions, and their
inattention and resisting behaviors, which indicate contrasting
inhibited attachment, were the same as well. This complete
absence of treatment effects for these behaviors suggests a
pivotal role for the stress factor. Most probably, the
precondition for activation of behaviors motivated by consolation
seeking--stress--was lacking or present in a very weak form for
most infants and the behavior ratings obtained do not reflect true
attachment responses.
The minimal response to separations and casual response to
reunions are best interpreted as consequences of familiarity with
brief separations from attachment figures and frequent entries
into new environments. Most of the irfants went to at least one
household other than their home on a daily basis and experienced
separations from attachment figures of far longer duration than
those entailed in the lahniatfiry procedure without upset. Manv
parents reperted that infants were accustomed tu regular visitF.
hurries ut re,Atives and friends. and %,isits with.:.ut 'hem .te
African Amer.
39
homes of relatives and friends of supplementary caretakers. As a
probable consequence, the strangeness of the laboratory playroom
was construed as novelty and aroused the infants' curiosity. Thus
for most infants with prior experiences like those in this study
something much more disconcerting than brief separations irom
attachment figures while remaining in a pleasant room with toys
and a friendly stranger is needed to effectively trigger
attachment responses to attachment figures. The results of the
analysis of variance for repeated measures comparisuns for the
proximity seeking and contact maintaining variables suggest that
the procedure Usti is effective in inducing uninhibited attachment
responses in a minority African American infants. Nonetheless,
pending the invention of a procedure that more effectively induces
stress in most infants like those studied two things remain'
unknown: Whether or not nonpathological proximity promoting
behaviors are the normative response of distressed African
American infants toward their attachment figures, and whether or
not the monotropy thesis applies to infants' di,:tressed state
responses toward attachment figures.
There are, of course, alternative explanations for the
reunion behavior outcomes. There is the possibility that the
measures used were too insensitive to detect meaningful variation
in infant reunion responses. Thls, however, seems improbable
given the fact that they are est*entially the same as Ainsworth's
rzting scales. There :s also !he rvifitility thist reurlons
4 0
African Amer.
40
that were supposed to tap behavior ensuing from a stressed state
did so for many of the infants, but diversity of individual
differences made normative responses undetectable. Along with the
repeated measures analysis of variance results, the
intercorrelations of variables provide some information pertinent
to this last possibility.
intercorrelations of dependent measures revealed that there
were individual differences among infants in overall response to
the laboratory procedure and, most probably, associated individual
differences in response to the episode changes in the procedure.
This is the most likely explanation for the differences in
variances of scores for episodes used in each of the planned
comparisons. However, indivioual differences were not a broadly
varied assortment; the.), reflected a single bipolar dimension of
extraversion verses wariness and staying close to attachment
figures. The majority if infants exhibited a gregariousness that
is encouraged by African American culture (Young, 1970). They
engAgod in exploratory and sociable behaviors and were inattentive
to their attachment figures' returns to the room; they also
exhibited little distress, proximity seeking or contact
maintaining behavior. Others who were distressed engaged in
proximity seeking and contact maintaining behavior instead of
exploratory and sociable behavior. Overall. findings from the
intercorrelations of dependent measures complement those revealing
substantial although incomplete support ot the research
African Amer.
41
hypotheses They also provide support for the tentative
inference that culturally sanctioned patterns of caregiving do not
produce pathological infant-attachment figure relationships for
most infants while they do produce infant traits censonant with
African American socialization objectives.
There are many issues that the findings presented impiy but
cannot address that future studies could fruitfully examine for a
fuller understanding of attachment phenomena in African American
infants--factors such as infant behavior toward all identifiable
attachment figures, configurations of caregiving arrangements,
stability of caregiving arrangements, histories of caregiver-
infant interactions, post-natal health, temperament and sex
differences are among them. Overall, more study of attachment
phenomena in African American children is needed, especially
research that uses culLurally based patterns of child care as a
point of departure. The issues that were examined and the
findings reported only liegin to fill the void in information on
African American children's attachments. Viewed in a broad
perspective, they highlight the need for exacting study of
attachment relationships of infants from diverse cultural groups
from a variety of perspectives including the value bases of the
cultures studied.
African Amer.
42
References
Ainsworth, M.D. (1967). inbulsaartaliland.k. Baltimore: Johns
Hopkins Press.
Ainsworth, M.D. (1982). Attachment: Retrospect and Prospect. In
C.M. Parkes & J. Stevenson-Hinde (Eds.), The Place of atttachment
inhuman behavior, (pp. 3-30). New York: Basic Books.
Ainsworth, M.D., Blehar, M.C., Waters, E., & Wall, S. (1978).
patterns ot,attechmenti A psychological st4pv qj the still:um
situation. Hillsdale NJ: Erlbaum.
Ainsworth, M.D., & Wittig, B.A. (1989). Attachment and
exploratory behavior of one-year-olds in a strange situation.
In B.M. Foss (Ed.), pidualmanttualatantjudausu.11 (pp. 111-
138). London: Meuthen.
Barnes, E.J. (1960). The Black community as the source of
positive self-concept for Black children: A theoretical
perspective. In R.L. Jones (Ed.), Black meYcholoss, (2nd ed.).
(pp. 106-130). New York: Harper & Row.
Beckett, J.0. (1E76). Working wives: A racial comparison.
ioial Work, al(6), 463-471.
Berry, J.W. (1979). Introduction to Methodology. In H.C.
Triandis & J.W. Berry (Eds.), Handbook- oi cross-cultural
psychology: Methodology, Vol.2. (pp. 1-261. Boston: Allyn &
Bacon.
Billingsley, A. (1968). Black families in wt.ite Ametica.
Englewood Clitts. NJ: Prentice-hail.
African Amer.
43
Blau, P., & Duncan, O. (1987). ThLkatiguussugatignai
structure. New York: Wiley & Sons.
Bowlby, J. f1951). Maternal care and mental health. pulletin
PI the World hpalth Ortanizatton, 355-534.
Bowlby, J. (1958). The nature of the child's tie to his mother.
International Journal qf twhoanalysis, aft, 350-373.
Bowlby, J. (1969). Attachment and lose Vol. I: Attachment. New
York: Basic Books.
Elowlby, J. (1973). Attachment and lose VoLu_LL: Separation.
Amiety and Anger. New York: Basic Books.
Bowlby, J. (1982). Wachment end loss Vol, Is Attachment (2nd
ed.). New York: basic Books.
B.M., Hershey, L., Lipton, E.L., Richmond, J.B., Stern.
G.A., Eddy, E., Drachman, R., id Rothman, A. (1963). Mother-
infant interaction in monomatric and polymatric families.
American Journi4 of Orthopsychiatry. al, 653-664.
Campbeil, D.T., 5 Stanley, J.C. (1966). ilurimentaj
and quasi-experimental desians lor research. Chicago: Rand
McNally.
Children's Defense Fund. (1985). Black and white children:
facts. Washington DC: Author.
Clarke-Stewart, K.A. (1973). Interactions between mothers and
their young children: Characteristics and consequences.
Mono fa hs o the Socie*.v for Research ,n Child Development._
36(6-7, Serial No. 1531.
4 4
African Amer.
44
Crittenden, P.M. (1988). Ralationships at risk. In J. Belsky &
T. Nezworski (Eds.), Clinical Lmolicati.ns of attachment, (pp.
138-178). Hillsdale, NJ: Erlbaum.
DeVore, I., & Konner, M.J. (1974). Infancy in hunter-gatherer
life: An ethological perspective. In N.F. White (Ed.),
Etholoty _ancLksYchiatry (113-141). Toronto: University of
Toronto Press.
Fox, N. (1977). Attachment of kibbutz infants to mother
and metapelet. Child Development, ill 1228-1239.
Goosens, F. A. (1987). Maternal employment and day-care: Effects
of attachment. In L.W.C. Tavecchio & M.H. van lJzendoorn
(Eds.), Attachment in social netwolks: Contributions to the
Bowlby-Ainsworth attachment theory (pp. 135-174). New Yorks
Elsevier Science Publishing.
Hansen, J.F. (1980). Children's attachment to mother _figures and
the social development of black infants and Young children.
Unpublished manuscript, University of California, Berkeley,
Psychology Department.
Hayghe, H. (1986). Rise in mothers' labor force activity includes
those with infants. Monthly Labor Review, 109(2), 43-45.
Hock, E. (1980). Working and non-working mothers and their
infants: A comparative study of maternal caregiving character-
istics and infant social behavior. Merrill-Falmer Quarterly,
26, 79-101.
Jackson. J.F. lin press). An experimental procedure and
African Amer.
45
scales for assessing attachment relationships in black infants--
alternatives to Ainsworth methods. In R.L. Jones (Ed.),
Handbook of tests and metstures cit black_looulations.
Kermoian, R., & Leiderman, P.N. (1966). Infant attachment to
mother and child caretaker in an east African community.
Llaternational Journal of Behavioral Develooment, gpo, 455-469.
Kotelchuck, M. (1976). The infant's relationship to the father:
Experimental evidence. In M.E. Lamb (Ed.), The role of the
father in child Oveloppent (pp. 329-344). New York: Wiley &
Sons.
Kratochwill, T.R. (1976). Foundations of time-series research.
In T.R. Kratochwill (Ed.). Angie subieql relparch; Strategies
for evaluating change (pp. 1-100). New York: Academic Press.
Kuhn, D. (Ed.). (199,). Special topic: Cross-cultural valWity
of attachment theory. Human Develooment,,
Lee, R.B. (1979). The !Kunt Sam Men women and wekik in a
foraging society. Cambridge: Cambridge University Press.
Lee, R.B., & DeVore, 1. (1976). Kalahari hunter gatherers:
Studies of the !Kung San and their neighbors. Cambridge, MA:
Harvard University Press.
Leiderman, P.N., & Leiderman, G.F. (1974). Affective and
cognitive consequences of polymatric infant care in the east
African highlands. In A.D. Pick (Ed.) Minnesota Symposium on
Child Psychology, Vol. VIII. (pp. 81-110). Minneapolis:
University of Minnesota Press.
4 6
African Amer.
46
LeVine, R.A. (1990). infant environments in psychoanalysis: A
cross-cultural view. In J.W. Stigler, R.A. Shweder, & G. Herdt
(Eds.), Cultural PSYCh0101Y: ESSAYS nm comparative human
develment. (pp. 454-474). New York: Cambridge University
Prete.
LeVine, R.A., & Miller, P.M. (1990).. Commentary. haan
Pevelopment 11(1), 73-60.
Lyons-Ruth, K., Connell, D.B., & Grunebaum, H.U. (1990). Infants
at social risk: Maternal depression and family support services
as mediators of infant development and security of attachment.
Child DeveloPment, W1), 65-96.
Marascuilo, L.A. (1971). Statistical methpds fqr behavioral.
sctknce research. New York: McGraw-Hill.
Marascuilo, L.A., & McSweeney, M. (1977). Nonparametric And
Oistribmitionlyee methods tor the social sciences. Monterey CA:
Brooks/Cole.
Marvin, R.S., VanDevender, T.L., lwanaga, M.I., LeVine, S., &
LeVine, R.A. (1977), lnfant-caregiver attachment among the
Hausa of Nigeria. In H. McGurk (Ed.), Ecolotical factors in
LWILIAJMUSIELLII (Pp. 247-260). Amsterdam: North-Holland.
Meisenheimer, J.R. %1990). Black college graduates in the labui
market, 1979 and 1989. Moattlx_kabor Review, 113(11), 13-21,
Nahemow, N.R. 11984). Grandparenthood in transition. In K.A.
McCluskey & H.W. Reese (Eds.), Life-span developmental psycheic-
;): Historical ettf-ctt--.. (pp. ;:49-277). Orlando: Acaaemic.
4. 7
African Amer.
47
O'Connell, M., & Rogers, C.C. (1983). Child care arrantements of
working mgthers: June 1982 (Current Population Reports Series
P-23, No. 129). Washington DC: U.S. Government Printing
Office.
Ogbu, J.U. (1981). Origins of human competence: A cultural-
ecological perspective. Child Development, kg(2), 413-429.
Presser, H.B. (1986). Shift work among American women and child
care. lournaL of Marriage and the family, 11, 551-563.
Presser, H.B. (1987). Work shifts of tull-time dual-earner
couples: Patterns and contrasts by sex of spouse. Demography,
24(1), 99-112.
Presser, H.B. (1989). Some economic complexities of child care
provided by grandmothers. lournal of Aarljage and the Family,
al, 581-591.
Reed, G., & Leiderman, P.H. (1981). Age-related changes in
attachment behavior in polymatrically reared infants: The
Kenyan Gusii, In T.M. Field, A.M. Sostek, P. Vietze & P.H.
Leiderman (Eds.), Culture and early interactions (pp. :15-2:A).
Hillsdale NJ: Erlbaum.
Research and Clinical Center for Child Development, (199)). is
the Ainsworth strange situation applicable to the study of
attachment in Japanese infants? (Annual Report Occasional Paper':
No. 2). Sapporo, Japan: author.
Sagi, A., Lamb, M.E., Lewkowic.I, K.S., Shoham, k., Dvir, R., &
Estes, D. (1965). In I. Bretherton & E. Waters iEds.1, Growing
f7Pf.
.g
African Amer.
48
points of attachment theory and research (pp. 257-275).
Monographs of the SocietY for Research.in Child Development.
k9.(1-2, Serial No. 209).
Scanzoni, J. (1975). Sex roles, economic factors, and marital
solidarity in black and white marriages. Journal of Marriage
and the Family, az, 130-144.
Shimkin, D., Louie, G., & Frets, D. (1978). The black extended
family: A basic rural isstitution and a mechanism of urban
adaptation. In D. Shiskin, G. Louie & D. Frate (Eds.), /ht
extended famiLY in black Societies, (pp. 25-147). The Hague:
Mouton.
Smith, P.K., & Noble, R. (1987). Factors affecting the
development of caregiver-infant relationships. In L.W.C.
Tavecchio & N.H. van lJzendoorn (Eds.), AttachmeAtinfipcial
nekidotyss Contributions to the Bowlby-Ainsworth attachment
theory, (pp. 93-131). New Yorks Elsevier Science Publishing.
Stack, C.B. (1970). The kindred of Viola Jackson: Residence of
an urban black American family. In N.E. Whitten 8 J.F. Szwed
(Eds.), Afro-american anthropojoiy: Contemporary perspectives
(pp. .W3-311). New York: Free Press.
Stack, C.B. (1974). All our kin: StrateRies for survival in a
black community. New York: Harper & Row.
Stack, C.B. (1975). Who raises black children: Transactions of
child givers and child receivers. In T.R. Williams (Ed.),
Socialization and communication in primacy groups (wor!d
African Amer.
49
enthrobolosv) (pp. 183-205). The Hagues Mouton.
Stinson, J.F. (1990. ) Multiple jobholding up sharply in the
1960's. Monthly Labor Review, 111(7), 3-10.
Tavecchio, L.W.C., & van 1.1zendoorn, M.H. (1987). Perceived
security and extension of the child's eearing context: A
parent-report approach. In L. W.C. Tavecchio & M.H. van
IJzendoorn (Eds.), Attachment in social networks: Cqptributions
to the Bowlbv-Ainsvorth attahmept theory, (pp. 36-92). New
York: Elsevier Science Publishing.
Tronick, E.Z., Morelli, G.A., & Winn, S. (1987). Multiple
caretaking of Efe (Pgymy) Infants. American Anthropgjoeist,
98-106.
Tronick, E.Z., Winn, S., & Morelli, G.A. (1985). Multiple
caretaking in the context of human evolution: Why don't the Efe
know the western prescription for child care? In M. Betts & T.
Field (Eds.), Theiosychobioloey of attachment and SepqratIon
U.S. Bureau of the Census (1974). 6tajist1cal Abstracts of the
United States: 1974 (95th ed.), (p. 336). Washington DC: U.S.
Government Printing Office.
van ljzendoorn, N.H. (1990). Developments in cross-cultural
research on attachment: Some methodological notes. Human
Development, 33(1), 3-9.
Waters, E.. Vaughn, B.E.. & Egeland. B.R. (1960). Individual
differences in infant-mother attachment relationships at age
one: Antecedents in neonatal behavior in an urban, economical.v
African Amer.
50
disadvantaged sample. Child Devaloomeol, §1,(1), 208-216.
Werner, E.E. (1984). Dild caret Kith. kinh and hired hands.
Baltimore: University Park Press.
White, M.M. (1987). We are family!: Kinship and solidarity in
the Black community. In G. Gay and W.L. Baber (Eds.),
Expressively neck: The cultuTal baste of ethnic identity, (pp.
17-34). New Yorks Praeger.
Winn, S., Tronick, E.Z., Morelli, G.A. (1989). The infant and the
group: A look at Efe Caretaking Practices in Zaire. In J.K.
Nugent, B.M. Lester, & T.B. Brazelton (Eds.), The cultural
context of infancy. Vol, IL BiolotY. culture. and infknt
develop:Ant, (pp. 87109).
Young, V.H. (1970). Family and childhood in a southern negro
community. American Anthropologist, 21(2), 259-288.
African Amer.
51
Table 1
Admjtj Present ith tnlant Utt SMOCASSiVe Boisode of
Elperiment
Episode
1
2
Mother Departs First
mother, adult M2
adult *2
Adult 02 Departs First
adult *2, mother
mother
3 adult 02, stranger mother, stranger
4 stranger stranger
5 stranger, mother stranger, adult M2
mother adult 112
7 mother, adult M2 adult *2, mother
8 mother adult *2
9 mother, stranger adult 42, stranger
10 stranger stranger
11 stranger, adult *2 stranger, mother
12 adult *2 mother
13 adult *2, mother mother, adult N2
Note. The term "adult Ir.." reters to any one of the
biological fathers or 5 female relatives who
participated as second attachment figures.
African Amer.
Table 2
actrelatiOnA 0 infante' Mother inftmenmed and_Setcond
Attachment Fitute Lnfluenced Behaviozs
cZALOL4Zer:0 Variable Pair.
Exploration .42*
Influenced by mother 3.9 1.17
Influenced by adult *2 3.81 1.31
Sociability .38*
Influenced by mother 4.05 1.63
Influenced by adult 02 4.21 1.69
Distress .65"e
Influenced by mother 1.84 1.06
Influenced by adult 02 1.71 .97
Proximity Seeking .661"
Influenced by mother 2.87 1.65
Influenced by adult *2 2.63 1.25
Contact Maintaining .340
Influenced by mother 1.82 1.26
Influenced by adult $2. 1.95
Inattention .42°'
Influenced by moth,ar 1. 95 1.09
Influenced by adult 2..26 1. -38
(table continues.
52
Variable Pair.
African Amer.
Resisting
Influenced by mother 1.11 .26
Influenced by adult 02 1.11 .26
-.14
53
Roll. The term "adult 02" refers to any adult from the
set of 37 second attachment figures.
aft 37 for all tests
°E. < .05. "B. .01. E (.001.
.
African Amer
Table 3
Matched Pair t-lests_ foT Mather influenced and Rtikreal
altaraIMILLUARTA-LIALYAIISELigliLYIALL
Difference Score
Variables d of d U. f S.
Exploration .13 1.35
Sociability -.18 1.85
Distress .14 .86
Proximity Seeking .24 1.29
Contact Maintaining -.14 1.46
Inattention -.28 1.35
Resisting -.01 .42
.se
-.53
.96
1.14
-.56
-1.28
-.10
54
Note. For all variables second attachment figurP
influenced scores were subtracted from mother influenced
scores to produce difference scores.
i'c 'N 37,/(3i a .05 for each test;
Af r lain Water
Axpit.eicr.0
Episode ISO its
Exploration
55
Contrast Coefficients
H vs. L. H vs. M M vs. L
1 5.42 1.88 10.28 1 2
2 4.92 2.14 9.35 1 2
3 4.25 2.65 7.89 -3.5 1
4 1.85 2.12 3.47 -3.5 -2
5 3.32 3.85 6.42 -3.5 1
6 4.03 3.03 7.69 1 2
7 4.20 4.22 8.03 1 2
8 4.20 3.91 8.00 1 2
9 3.08 3.47 5.99 -3.5 1
10 1.87 2.45 3.62 -3.5 --..
11 2.87 3.68 5.31 -3.5 1
12 3.89 4.54 7.46 1 2
13 3.64 4.06 7.16 1 2
(table continue)
Episode
African Aber.
58
Contrast Coefficients
Sociability
3 3.81 4.82 3.77 1
4 4.88 4.34 3.81 -2
5 3.32 4.23 2.58 i
9 4.14 5.12 3.65 1
10 4.32 4.34 3.45 -2
11 5.05 4.39 4.38 1
Distress
2 1.06 .11 1.96 1
4 1.92 2.19 2.84 -1
8 1.70 1.49 2.49 1
10 2.32 3.23 2.97 -1
Proximity Seeking
5 2.78 3.45 2.54 1
7 2.27 3.42 2.32 -1
11 3.41 4.64 2.85 1
13 2.35 3.12 2.39 1
(table continuef)
Episodo7 .
5
6
7
a
11
12
13
5
7
11
13
AfriOan Amer.
57
Contrast Cofficients
Contact Maintaining
1.84 3.03 4.05 2.5
1.35 1.01 3.54 .1
1.24 .52 3.34 .1
1.97 2.84 4.05 .1
2.78 5.23 4.88 2.5
2.19 4.66 4.16 -1
1.65 1.85 3.86 .1
,nattention
2.16 3.03 2.55 1
2.03 2.14 2.45 .1
1.92 2.02 2.36
2.22 2.34 2.82 .1
(table continues)
Episode
African Amer.
SO
ti B! Contrast Coefficients
Resisting
5 1.03 .03
6 1.08 .13
7 1.19 .27
8 1.05 .11
11 1.11 .18
12 1.24 .36
13 1.08 .13
Kara. N is 37 for all comparisons.
'Mean of within subject ranks.
3.89 2.5
3.91 -1
4.27 -1
3.81 -1
4.00 2.5
4.27 .1
3.88 -1
5 5
Table Li
c_mcasiscno
Variable Var.'
Uttet
V/511 xploration:
%II a 33.38 - (13.38)1/1(8.39)11, 8.39
24.14'A
hi. 28.75 - (15.99)1,8(22.89)402 22.89
7.81'
11.43 - (9.24)11(1.14)1#2 1.14
8.199
Sociability:
u .1 , (11.07)111 (1,14)i /a 1.14
African Amor.
59
.0.---x"a"m"""m"-
11-Critical Value
Dist544-11,e-r r
0 (6.25)4'1(.1858)1'8 .1858
- .08.
A
id g -1.16 + (6.25)1,2(.18)1,2 .le
= - .10.
Proximity Seeking:
* .68 - (6.25)1,2(.18)117 .18
Statistic di 1-ct.
13.38 8 .90
15.89 10 .90
9.24 5 .90
11.07 5 .95
6.25 3 .90
6.25 3 .90
6.25 3 .90
(table continues)
Variable
Contact Maintaining:
3.38 - (10.64)1'1(2.21)10*
Inattention:
yis - .17 * (7.61)1f1(.16)111
Resisting:
yis - .48 (12.59)112(2.21)11*
timo% nwrZ67,--yA
eVar.IJ k(k-1)/12 al.1/n (Marascuilo40
*Contrast value corrected for ties.
4.E.< .05, one-tailed.
African Amer.
60
I-Critical Value
Vardi Statistic dt 1
2.21 10.84 8 .90
.18 7.81 3 .95
2.21 12.59 6 .95
Mcsweeney, 1977, pp. 362-368).
_ Table 6
African Amer.
61
Vonfidence lntervaLs And Statistics for Comparisons Using Repeated,
twilialsLitnajyjaiLsi_ibuianst
Variable
S.
4;440._ d-vow f jC.0 Iseutr. 1
- Critical Value
Var.* Statistic dfs die
Sociability:v.
4) -1.88 . (3.32)(1.11)1'2 1.11 3.32 5 180 .95
Proximity Seeking:
Y a 1.57 - (2.55)(.31)lis
a .14*
Contact Maintaining:
m 3.15 - (3.26)(.81)1,2
a .22*
Inattention:
y- a - .17 (2.87)(.20)1/1
. 31 2.55 3 108 .90
. 81 3.26 6 218 .90
. 20 2.87 3 108 .95
Resisting:
- .29 (3.55)(.05)1/: .05 3.55 6 Z16 .95
l4xCcc',.r. GC* .(^
tri = EvtF.1..:(1.X))"4* plartmt 0,2:n ft:A (Marascuilo, 1971, pp.get
458-467).
op< .05, one-tailed.
African Amer.
62
. Table 7
Variable 2. 3. 4. S. e. 7.
1. Exploration .26
pa.06
2. Sociability --
3. Inattention
4. Resisting
S. Proximity
Seeking
6. Contact
Maintaining
7. Distress
.10 .22 -.32* -.46** -;.39*
.32* .24 -.33* -.38* -.496*
.17 -.35* -.22ainliNes
-.08 -.21 -.13
IND .609" .43"
Note. N 2 37 for all correlations
ge. < .05. " < .01. wa < .001.
6:3
IND .59".
1 2 3 4 5 6 7 8 9 10 11 12 13A, A, AI A2 A2 Al A2 A2 AI AI A,A2 S S $ AI S S S A2
ADULTS IN ROOM BY EPISODE
A, 1 Mother or Second Attachment Figure,A2 j Depending on Order of Departure
S Stranger
4
African Amer.
63
Figure Caption
Figure 1. Level of infant explorat4on in means of within subject ranks
of exploration scores as an effect of adults in the room in each episode
of the experimental session.
65